go-ethereum/p2p/discover/udp.go

756 lines
22 KiB
Go
Raw Normal View History

2015-07-06 19:54:22 -05:00
// Copyright 2015 The go-ethereum Authors
// This file is part of the go-ethereum library.
2015-07-06 19:54:22 -05:00
//
// The go-ethereum library is free software: you can redistribute it and/or modify
2015-07-06 19:54:22 -05:00
// it under the terms of the GNU Lesser General Public License as published by
// the Free Software Foundation, either version 3 of the License, or
// (at your option) any later version.
//
// The go-ethereum library is distributed in the hope that it will be useful,
2015-07-06 19:54:22 -05:00
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
2015-07-06 19:54:22 -05:00
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public License
// along with the go-ethereum library. If not, see <http://www.gnu.org/licenses/>.
2015-07-06 19:54:22 -05:00
package discover
import (
"bytes"
"container/list"
"crypto/ecdsa"
"errors"
"fmt"
"net"
"sync"
"time"
"github.com/ethereum/go-ethereum/crypto"
"github.com/ethereum/go-ethereum/log"
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
"github.com/ethereum/go-ethereum/p2p/enode"
"github.com/ethereum/go-ethereum/p2p/netutil"
"github.com/ethereum/go-ethereum/rlp"
)
// Errors
var (
errPacketTooSmall = errors.New("too small")
errBadHash = errors.New("bad hash")
errExpired = errors.New("expired")
errUnsolicitedReply = errors.New("unsolicited reply")
errUnknownNode = errors.New("unknown node")
errTimeout = errors.New("RPC timeout")
errClockWarp = errors.New("reply deadline too far in the future")
errClosed = errors.New("socket closed")
)
// Timeouts
const (
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
respTimeout = 500 * time.Millisecond
expiration = 20 * time.Second
bondExpiration = 24 * time.Hour
ntpFailureThreshold = 32 // Continuous timeouts after which to check NTP
ntpWarningCooldown = 10 * time.Minute // Minimum amount of time to pass before repeating NTP warning
driftThreshold = 10 * time.Second // Allowed clock drift before warning user
)
// RPC packet types
const (
pingPacket = iota + 1 // zero is 'reserved'
pongPacket
findnodePacket
neighborsPacket
)
// RPC request structures
type (
ping struct {
senderKey *ecdsa.PublicKey // filled in by preverify
Version uint
From, To rpcEndpoint
Expiration uint64
2015-12-22 18:52:40 -06:00
// Ignore additional fields (for forward compatibility).
Rest []rlp.RawValue `rlp:"tail"`
}
// pong is the reply to ping.
pong struct {
// This field should mirror the UDP envelope address
// of the ping packet, which provides a way to discover the
// the external address (after NAT).
To rpcEndpoint
ReplyTok []byte // This contains the hash of the ping packet.
Expiration uint64 // Absolute timestamp at which the packet becomes invalid.
2015-12-22 18:52:40 -06:00
// Ignore additional fields (for forward compatibility).
Rest []rlp.RawValue `rlp:"tail"`
}
// findnode is a query for nodes close to the given target.
findnode struct {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
Target encPubkey
Expiration uint64
2015-12-22 18:52:40 -06:00
// Ignore additional fields (for forward compatibility).
Rest []rlp.RawValue `rlp:"tail"`
}
// reply to findnode
neighbors struct {
2015-04-23 05:11:21 -05:00
Nodes []rpcNode
Expiration uint64
2015-12-22 18:52:40 -06:00
// Ignore additional fields (for forward compatibility).
Rest []rlp.RawValue `rlp:"tail"`
}
2015-04-23 05:11:21 -05:00
rpcNode struct {
IP net.IP // len 4 for IPv4 or 16 for IPv6
UDP uint16 // for discovery protocol
TCP uint16 // for RLPx protocol
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
ID encPubkey
2015-04-23 05:11:21 -05:00
}
rpcEndpoint struct {
IP net.IP // len 4 for IPv4 or 16 for IPv6
UDP uint16 // for discovery protocol
TCP uint16 // for RLPx protocol
}
)
func makeEndpoint(addr *net.UDPAddr, tcpPort uint16) rpcEndpoint {
ip := net.IP{}
if ip4 := addr.IP.To4(); ip4 != nil {
ip = ip4
} else if ip6 := addr.IP.To16(); ip6 != nil {
ip = ip6
}
return rpcEndpoint{IP: ip, UDP: uint16(addr.Port), TCP: tcpPort}
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func (t *udp) nodeFromRPC(sender *net.UDPAddr, rn rpcNode) (*node, error) {
if rn.UDP <= 1024 {
return nil, errors.New("low port")
}
if err := netutil.CheckRelayIP(sender.IP, rn.IP); err != nil {
return nil, err
}
if t.netrestrict != nil && !t.netrestrict.Contains(rn.IP) {
return nil, errors.New("not contained in netrestrict whitelist")
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
key, err := decodePubkey(rn.ID)
if err != nil {
return nil, err
}
n := wrapNode(enode.NewV4(key, rn.IP, int(rn.TCP), int(rn.UDP)))
err = n.ValidateComplete()
return n, err
2015-04-23 05:11:21 -05:00
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func nodeToRPC(n *node) rpcNode {
var key ecdsa.PublicKey
var ekey encPubkey
if err := n.Load((*enode.Secp256k1)(&key)); err == nil {
ekey = encodePubkey(&key)
}
return rpcNode{ID: ekey, IP: n.IP(), UDP: uint16(n.UDP()), TCP: uint16(n.TCP())}
}
// packet is implemented by all protocol messages.
type packet interface {
// preverify checks whether the packet is valid and should be handled at all.
preverify(t *udp, from *net.UDPAddr, fromID enode.ID, fromKey encPubkey) error
// handle handles the packet.
handle(t *udp, from *net.UDPAddr, fromID enode.ID, mac []byte)
// name returns the name of the packet for logging purposes.
name() string
}
type conn interface {
ReadFromUDP(b []byte) (n int, addr *net.UDPAddr, err error)
WriteToUDP(b []byte, addr *net.UDPAddr) (n int, err error)
Close() error
LocalAddr() net.Addr
}
// udp implements the discovery v4 UDP wire protocol.
type udp struct {
conn conn
netrestrict *netutil.Netlist
priv *ecdsa.PrivateKey
localNode *enode.LocalNode
db *enode.DB
tab *Table
wg sync.WaitGroup
addReplyMatcher chan *replyMatcher
gotreply chan reply
closing chan struct{}
}
// pending represents a pending reply.
//
// Some implementations of the protocol wish to send more than one
// reply packet to findnode. In general, any neighbors packet cannot
// be matched up with a specific findnode packet.
//
// Our implementation handles this by storing a callback function for
// each pending reply. Incoming packets from a node are dispatched
// to all callback functions for that node.
type replyMatcher struct {
// these fields must match in the reply.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
from enode.ID
ip net.IP
ptype byte
// time when the request must complete
deadline time.Time
// callback is called when a matching reply arrives. If it returns matched == true, the
// reply was acceptable. The second return value indicates whether the callback should
// be removed from the pending reply queue. If it returns false, the reply is considered
// incomplete and the callback will be invoked again for the next matching reply.
callback replyMatchFunc
// errc receives nil when the callback indicates completion or an
// error if no further reply is received within the timeout.
errc chan<- error
}
type replyMatchFunc func(interface{}) (matched bool, requestDone bool)
type reply struct {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
from enode.ID
ip net.IP
ptype byte
data packet
// loop indicates whether there was
// a matching request by sending on this channel.
matched chan<- bool
}
// ReadPacket is sent to the unhandled channel when it could not be processed
type ReadPacket struct {
Data []byte
Addr *net.UDPAddr
}
// Config holds Table-related settings.
type Config struct {
// These settings are required and configure the UDP listener:
PrivateKey *ecdsa.PrivateKey
// These settings are optional:
NetRestrict *netutil.Netlist // network whitelist
Bootnodes []*enode.Node // list of bootstrap nodes
Unhandled chan<- ReadPacket // unhandled packets are sent on this channel
}
// ListenUDP returns a new table that listens for UDP packets on laddr.
func ListenUDP(c conn, ln *enode.LocalNode, cfg Config) (*Table, error) {
tab, _, err := newUDP(c, ln, cfg)
if err != nil {
return nil, err
}
return tab, nil
}
func newUDP(c conn, ln *enode.LocalNode, cfg Config) (*Table, *udp, error) {
udp := &udp{
conn: c,
priv: cfg.PrivateKey,
netrestrict: cfg.NetRestrict,
localNode: ln,
db: ln.Database(),
closing: make(chan struct{}),
gotreply: make(chan reply),
addReplyMatcher: make(chan *replyMatcher),
}
tab, err := newTable(udp, ln.Database(), cfg.Bootnodes)
if err != nil {
return nil, nil, err
}
udp.tab = tab
udp.wg.Add(2)
go udp.loop()
go udp.readLoop(cfg.Unhandled)
return udp.tab, udp, nil
}
func (t *udp) self() *enode.Node {
return t.localNode.Node()
}
func (t *udp) close() {
close(t.closing)
t.conn.Close()
t.wg.Wait()
}
func (t *udp) ourEndpoint() rpcEndpoint {
n := t.self()
a := &net.UDPAddr{IP: n.IP(), Port: n.UDP()}
return makeEndpoint(a, uint16(n.TCP()))
}
// ping sends a ping message to the given node and waits for a reply.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func (t *udp) ping(toid enode.ID, toaddr *net.UDPAddr) error {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
return <-t.sendPing(toid, toaddr, nil)
}
// sendPing sends a ping message to the given node and invokes the callback
// when the reply arrives.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func (t *udp) sendPing(toid enode.ID, toaddr *net.UDPAddr, callback func()) <-chan error {
req := &ping{
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
Version: 4,
From: t.ourEndpoint(),
To: makeEndpoint(toaddr, 0), // TODO: maybe use known TCP port from DB
Expiration: uint64(time.Now().Add(expiration).Unix()),
}
packet, hash, err := encodePacket(t.priv, pingPacket, req)
if err != nil {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
errc := make(chan error, 1)
errc <- err
return errc
}
// Add a matcher for the reply to the pending reply queue. Pongs are matched if they
// reference the ping we're about to send.
errc := t.pending(toid, toaddr.IP, pongPacket, func(p interface{}) (matched bool, requestDone bool) {
matched = bytes.Equal(p.(*pong).ReplyTok, hash)
if matched && callback != nil {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
callback()
}
return matched, matched
})
// Send the packet.
t.localNode.UDPContact(toaddr)
t.write(toaddr, toid, req.name(), packet)
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
return errc
}
// findnode sends a findnode request to the given node and waits until
// the node has sent up to k neighbors.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func (t *udp) findnode(toid enode.ID, toaddr *net.UDPAddr, target encPubkey) ([]*node, error) {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
// If we haven't seen a ping from the destination node for a while, it won't remember
// our endpoint proof and reject findnode. Solicit a ping first.
if time.Since(t.db.LastPingReceived(toid, toaddr.IP)) > bondExpiration {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
t.ping(toid, toaddr)
// Wait for them to ping back and process our pong.
time.Sleep(respTimeout)
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
}
// Add a matcher for 'neighbours' replies to the pending reply queue. The matcher is
// active until enough nodes have been received.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
nodes := make([]*node, 0, bucketSize)
nreceived := 0
errc := t.pending(toid, toaddr.IP, neighborsPacket, func(r interface{}) (matched bool, requestDone bool) {
reply := r.(*neighbors)
2015-04-23 05:11:21 -05:00
for _, rn := range reply.Nodes {
nreceived++
n, err := t.nodeFromRPC(toaddr, rn)
if err != nil {
log.Trace("Invalid neighbor node received", "ip", rn.IP, "addr", toaddr, "err", err)
continue
}
nodes = append(nodes, n)
}
return true, nreceived >= bucketSize
})
t.send(toaddr, toid, findnodePacket, &findnode{
Target: target,
Expiration: uint64(time.Now().Add(expiration).Unix()),
})
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
return nodes, <-errc
}
// pending adds a reply matcher to the pending reply queue.
// see the documentation of type replyMatcher for a detailed explanation.
func (t *udp) pending(id enode.ID, ip net.IP, ptype byte, callback replyMatchFunc) <-chan error {
ch := make(chan error, 1)
p := &replyMatcher{from: id, ip: ip, ptype: ptype, callback: callback, errc: ch}
select {
case t.addReplyMatcher <- p:
// loop will handle it
case <-t.closing:
ch <- errClosed
}
return ch
}
// handleReply dispatches a reply packet, invoking reply matchers. It returns
// whether any matcher considered the packet acceptable.
func (t *udp) handleReply(from enode.ID, fromIP net.IP, ptype byte, req packet) bool {
matched := make(chan bool, 1)
select {
case t.gotreply <- reply{from, fromIP, ptype, req, matched}:
// loop will handle it
return <-matched
case <-t.closing:
return false
}
}
// loop runs in its own goroutine. it keeps track of
// the refresh timer and the pending reply queue.
func (t *udp) loop() {
defer t.wg.Done()
var (
plist = list.New()
timeout = time.NewTimer(0)
nextTimeout *replyMatcher // head of plist when timeout was last reset
contTimeouts = 0 // number of continuous timeouts to do NTP checks
ntpWarnTime = time.Unix(0, 0)
)
<-timeout.C // ignore first timeout
defer timeout.Stop()
resetTimeout := func() {
if plist.Front() == nil || nextTimeout == plist.Front().Value {
return
}
// Start the timer so it fires when the next pending reply has expired.
now := time.Now()
for el := plist.Front(); el != nil; el = el.Next() {
nextTimeout = el.Value.(*replyMatcher)
if dist := nextTimeout.deadline.Sub(now); dist < 2*respTimeout {
timeout.Reset(dist)
return
}
// Remove pending replies whose deadline is too far in the
// future. These can occur if the system clock jumped
// backwards after the deadline was assigned.
nextTimeout.errc <- errClockWarp
plist.Remove(el)
}
nextTimeout = nil
timeout.Stop()
}
for {
resetTimeout()
select {
case <-t.closing:
for el := plist.Front(); el != nil; el = el.Next() {
el.Value.(*replyMatcher).errc <- errClosed
}
return
case p := <-t.addReplyMatcher:
p.deadline = time.Now().Add(respTimeout)
plist.PushBack(p)
case r := <-t.gotreply:
var matched bool // whether any replyMatcher considered the reply acceptable.
for el := plist.Front(); el != nil; el = el.Next() {
p := el.Value.(*replyMatcher)
if p.from == r.from && p.ptype == r.ptype && p.ip.Equal(r.ip) {
ok, requestDone := p.callback(r.data)
matched = matched || ok
// Remove the matcher if callback indicates that all replies have been received.
if requestDone {
p.errc <- nil
plist.Remove(el)
}
// Reset the continuous timeout counter (time drift detection)
contTimeouts = 0
}
}
r.matched <- matched
case now := <-timeout.C:
nextTimeout = nil
// Notify and remove callbacks whose deadline is in the past.
for el := plist.Front(); el != nil; el = el.Next() {
p := el.Value.(*replyMatcher)
if now.After(p.deadline) || now.Equal(p.deadline) {
p.errc <- errTimeout
plist.Remove(el)
contTimeouts++
}
}
// If we've accumulated too many timeouts, do an NTP time sync check
if contTimeouts > ntpFailureThreshold {
if time.Since(ntpWarnTime) >= ntpWarningCooldown {
ntpWarnTime = time.Now()
go checkClockDrift()
}
contTimeouts = 0
}
}
}
}
const (
macSize = 256 / 8
sigSize = 520 / 8
headSize = macSize + sigSize // space of packet frame data
)
var (
headSpace = make([]byte, headSize)
// Neighbors replies are sent across multiple packets to
// stay below the 1280 byte limit. We compute the maximum number
// of entries by stuffing a packet until it grows too large.
maxNeighbors int
)
func init() {
p := neighbors{Expiration: ^uint64(0)}
maxSizeNode := rpcNode{IP: make(net.IP, 16), UDP: ^uint16(0), TCP: ^uint16(0)}
for n := 0; ; n++ {
p.Nodes = append(p.Nodes, maxSizeNode)
size, _, err := rlp.EncodeToReader(p)
if err != nil {
// If this ever happens, it will be caught by the unit tests.
panic("cannot encode: " + err.Error())
}
if headSize+size+1 >= 1280 {
maxNeighbors = n
break
}
}
}
func (t *udp) send(toaddr *net.UDPAddr, toid enode.ID, ptype byte, req packet) ([]byte, error) {
packet, hash, err := encodePacket(t.priv, ptype, req)
if err != nil {
return hash, err
}
return hash, t.write(toaddr, toid, req.name(), packet)
}
func (t *udp) write(toaddr *net.UDPAddr, toid enode.ID, what string, packet []byte) error {
_, err := t.conn.WriteToUDP(packet, toaddr)
log.Trace(">> "+what, "id", toid, "addr", toaddr, "err", err)
return err
}
func encodePacket(priv *ecdsa.PrivateKey, ptype byte, req interface{}) (packet, hash []byte, err error) {
b := new(bytes.Buffer)
b.Write(headSpace)
b.WriteByte(ptype)
if err := rlp.Encode(b, req); err != nil {
log.Error("Can't encode discv4 packet", "err", err)
return nil, nil, err
}
packet = b.Bytes()
sig, err := crypto.Sign(crypto.Keccak256(packet[headSize:]), priv)
if err != nil {
log.Error("Can't sign discv4 packet", "err", err)
return nil, nil, err
}
copy(packet[macSize:], sig)
// add the hash to the front. Note: this doesn't protect the
// packet in any way. Our public key will be part of this hash in
// The future.
hash = crypto.Keccak256(packet[macSize:])
copy(packet, hash)
return packet, hash, nil
}
// readLoop runs in its own goroutine. it handles incoming UDP packets.
func (t *udp) readLoop(unhandled chan<- ReadPacket) {
defer t.wg.Done()
if unhandled != nil {
defer close(unhandled)
}
// Discovery packets are defined to be no larger than 1280 bytes.
// Packets larger than this size will be cut at the end and treated
// as invalid because their hash won't match.
UDP Interop. Limit datagrams to 1280bytes. We don't have a UDP which specifies any messages that will be 4KB. Aside from being implemented for months and a necessity for encryption and piggy-backing packets, 1280bytes is ideal, and, means this TODO can be completed! Why 1280 bytes? * It's less than the default MTU for most WAN/LAN networks. That means fewer fragmented datagrams (esp on well-connected networks). * Fragmented datagrams and dropped packets suck and add latency while OS waits for a dropped fragment to never arrive (blocking readLoop()) * Most of our packets are < 1280 bytes. * 1280 bytes is minimum datagram size and MTU for IPv6 -- on IPv6, a datagram < 1280bytes will *never* be fragmented. UDP datagrams are dropped. A lot! And fragmented datagrams are worse. If a datagram has a 30% chance of being dropped, then a fragmented datagram has a 60% chance of being dropped. More importantly, we have signed packets and can't do anything with a packet unless we receive the entire datagram because the signature can't be verified. The same is true when we have encrypted packets. So the solution here to picking an ideal buffer size for receiving datagrams is a number under 1400bytes. And the lower-bound value for IPv6 of 1280 bytes make's it a non-decision. On IPv4 most ISPs and 3g/4g/let networks have an MTU just over 1400 -- and *never* over 1500. Never -- that means packets over 1500 (in reality: ~1450) bytes are fragmented. And probably dropped a lot. Just to prove the point, here are pings sending non-fragmented packets over wifi/ISP, and a second set of pings via cell-phone tethering. It's important to note that, if *any* router between my system and the EC2 node has a lower MTU, the message would not go through: On wifi w/normal ISP: localhost:Debug $ ping -D -s 1450 52.6.250.242 PING 52.6.250.242 (52.6.250.242): 1450 data bytes 1458 bytes from 52.6.250.242: icmp_seq=0 ttl=42 time=104.831 ms 1458 bytes from 52.6.250.242: icmp_seq=1 ttl=42 time=119.004 ms ^C --- 52.6.250.242 ping statistics --- 2 packets transmitted, 2 packets received, 0.0% packet loss round-trip min/avg/max/stddev = 104.831/111.918/119.004/7.087 ms localhost:Debug $ ping -D -s 1480 52.6.250.242 PING 52.6.250.242 (52.6.250.242): 1480 data bytes ping: sendto: Message too long ping: sendto: Message too long Request timeout for icmp_seq 0 ping: sendto: Message too long Request timeout for icmp_seq 1 Tethering to O2: localhost:Debug $ ping -D -s 1480 52.6.250.242 PING 52.6.250.242 (52.6.250.242): 1480 data bytes ping: sendto: Message too long ping: sendto: Message too long Request timeout for icmp_seq 0 ^C --- 52.6.250.242 ping statistics --- 2 packets transmitted, 0 packets received, 100.0% packet loss localhost:Debug $ ping -D -s 1450 52.6.250.242 PING 52.6.250.242 (52.6.250.242): 1450 data bytes 1458 bytes from 52.6.250.242: icmp_seq=0 ttl=42 time=107.844 ms 1458 bytes from 52.6.250.242: icmp_seq=1 ttl=42 time=105.127 ms 1458 bytes from 52.6.250.242: icmp_seq=2 ttl=42 time=120.483 ms 1458 bytes from 52.6.250.242: icmp_seq=3 ttl=42 time=102.136 ms
2015-05-13 12:03:00 -05:00
buf := make([]byte, 1280)
for {
nbytes, from, err := t.conn.ReadFromUDP(buf)
if netutil.IsTemporaryError(err) {
// Ignore temporary read errors.
log.Debug("Temporary UDP read error", "err", err)
continue
} else if err != nil {
// Shut down the loop for permament errors.
log.Debug("UDP read error", "err", err)
return
}
if t.handlePacket(from, buf[:nbytes]) != nil && unhandled != nil {
select {
case unhandled <- ReadPacket{buf[:nbytes], from}:
default:
}
}
}
}
func (t *udp) handlePacket(from *net.UDPAddr, buf []byte) error {
packet, fromKey, hash, err := decodePacket(buf)
if err != nil {
log.Debug("Bad discv4 packet", "addr", from, "err", err)
return err
}
fromID := fromKey.id()
if err == nil {
err = packet.preverify(t, from, fromID, fromKey)
}
log.Trace("<< "+packet.name(), "id", fromID, "addr", from, "err", err)
if err == nil {
packet.handle(t, from, fromID, hash)
}
return err
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
func decodePacket(buf []byte) (packet, encPubkey, []byte, error) {
if len(buf) < headSize+1 {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
return nil, encPubkey{}, nil, errPacketTooSmall
}
hash, sig, sigdata := buf[:macSize], buf[macSize:headSize], buf[headSize:]
shouldhash := crypto.Keccak256(buf[macSize:])
if !bytes.Equal(hash, shouldhash) {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
return nil, encPubkey{}, nil, errBadHash
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
fromKey, err := recoverNodeKey(crypto.Keccak256(buf[headSize:]), sig)
if err != nil {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
return nil, fromKey, hash, err
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
var req packet
switch ptype := sigdata[0]; ptype {
case pingPacket:
req = new(ping)
case pongPacket:
req = new(pong)
case findnodePacket:
req = new(findnode)
case neighborsPacket:
req = new(neighbors)
default:
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
return nil, fromKey, hash, fmt.Errorf("unknown type: %d", ptype)
}
2015-12-22 18:52:40 -06:00
s := rlp.NewStream(bytes.NewReader(sigdata[1:]), 0)
err = s.Decode(req)
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
return req, fromKey, hash, err
}
// Packet Handlers
func (req *ping) preverify(t *udp, from *net.UDPAddr, fromID enode.ID, fromKey encPubkey) error {
if expired(req.Expiration) {
return errExpired
}
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
key, err := decodePubkey(fromKey)
if err != nil {
return errors.New("invalid public key")
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
}
req.senderKey = key
return nil
}
func (req *ping) handle(t *udp, from *net.UDPAddr, fromID enode.ID, mac []byte) {
// Reply.
t.send(from, fromID, pongPacket, &pong{
To: makeEndpoint(from, req.From.TCP),
ReplyTok: mac,
Expiration: uint64(time.Now().Add(expiration).Unix()),
})
// Ping back if our last pong on file is too far in the past.
n := wrapNode(enode.NewV4(req.senderKey, from.IP, int(req.From.TCP), from.Port))
if time.Since(t.db.LastPongReceived(n.ID(), from.IP)) > bondExpiration {
t.sendPing(fromID, from, func() {
t.tab.addVerifiedNode(n)
})
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
} else {
t.tab.addVerifiedNode(n)
}
// Update node database and endpoint predictor.
t.db.UpdateLastPingReceived(n.ID(), from.IP, time.Now())
t.localNode.UDPEndpointStatement(from, &net.UDPAddr{IP: req.To.IP, Port: int(req.To.UDP)})
}
func (req *ping) name() string { return "PING/v4" }
func (req *pong) preverify(t *udp, from *net.UDPAddr, fromID enode.ID, fromKey encPubkey) error {
if expired(req.Expiration) {
return errExpired
}
if !t.handleReply(fromID, from.IP, pongPacket, req) {
return errUnsolicitedReply
}
return nil
}
func (req *pong) handle(t *udp, from *net.UDPAddr, fromID enode.ID, mac []byte) {
t.localNode.UDPEndpointStatement(from, &net.UDPAddr{IP: req.To.IP, Port: int(req.To.UDP)})
t.db.UpdateLastPongReceived(fromID, from.IP, time.Now())
}
func (req *pong) name() string { return "PONG/v4" }
func (req *findnode) preverify(t *udp, from *net.UDPAddr, fromID enode.ID, fromKey encPubkey) error {
if expired(req.Expiration) {
return errExpired
}
if time.Since(t.db.LastPongReceived(fromID, from.IP)) > bondExpiration {
p2p/discover: move bond logic from table to transport (#17048) * p2p/discover: move bond logic from table to transport This commit moves node endpoint verification (bonding) from the table to the UDP transport implementation. Previously, adding a node to the table entailed pinging the node if needed. With this change, the ping-back logic is embedded in the packet handler at a lower level. It is easy to verify that the basic protocol is unchanged: we still require a valid pong reply from the node before findnode is accepted. The node database tracked the time of last ping sent to the node and time of last valid pong received from the node. Node endpoints are considered verified when a valid pong is received and the time of last pong was called 'bond time'. The time of last ping sent was unused. In this commit, the last ping database entry is repurposed to mean last ping _received_. This entry is now used to track whether the node needs to be pinged back. The other big change is how nodes are added to the table. We used to add nodes in Table.bond, which ran when a remote node pinged us or when we encountered the node in a neighbors reply. The transport now adds to the table directly after the endpoint is verified through ping. To ensure that the Table can't be filled just by pinging the node repeatedly, we retain the isInitDone check. During init, only nodes from neighbors replies are added. * p2p/discover: reduce findnode failure counter on success * p2p/discover: remove unused parameter of loadSeedNodes * p2p/discover: improve ping-back check and comments * p2p/discover: add neighbors reply nodes always, not just during init
2018-07-03 08:24:12 -05:00
// No endpoint proof pong exists, we don't process the packet. This prevents an
// attack vector where the discovery protocol could be used to amplify traffic in a
// DDOS attack. A malicious actor would send a findnode request with the IP address
// and UDP port of the target as the source address. The recipient of the findnode
// packet would then send a neighbors packet (which is a much bigger packet than
// findnode) to the victim.
return errUnknownNode
}
return nil
}
func (req *findnode) handle(t *udp, from *net.UDPAddr, fromID enode.ID, mac []byte) {
// Determine closest nodes.
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
target := enode.ID(crypto.Keccak256Hash(req.Target[:]))
t.tab.mutex.Lock()
closest := t.tab.closest(target, bucketSize).entries
t.tab.mutex.Unlock()
// Send neighbors in chunks with at most maxNeighbors per packet
// to stay below the 1280 byte limit.
p := neighbors{Expiration: uint64(time.Now().Add(expiration).Unix())}
var sent bool
for _, n := range closest {
all: new p2p node representation (#17643) Package p2p/enode provides a generalized representation of p2p nodes which can contain arbitrary information in key/value pairs. It is also the new home for the node database. The "v4" identity scheme is also moved here from p2p/enr to remove the dependency on Ethereum crypto from that package. Record signature handling is changed significantly. The identity scheme registry is removed and acceptable schemes must be passed to any method that needs identity. This means records must now be validated explicitly after decoding. The enode API is designed to make signature handling easy and safe: most APIs around the codebase work with enode.Node, which is a wrapper around a valid record. Going from enr.Record to enode.Node requires a valid signature. * p2p/discover: port to p2p/enode This ports the discovery code to the new node representation in p2p/enode. The wire protocol is unchanged, this can be considered a refactoring change. The Kademlia table can now deal with nodes using an arbitrary identity scheme. This requires a few incompatible API changes: - Table.Lookup is not available anymore. It used to take a public key as argument because v4 protocol requires one. Its replacement is LookupRandom. - Table.Resolve takes *enode.Node instead of NodeID. This is also for v4 protocol compatibility because nodes cannot be looked up by ID alone. - Types Node and NodeID are gone. Further commits in the series will be fixes all over the the codebase to deal with those removals. * p2p: port to p2p/enode and discovery changes This adapts package p2p to the changes in p2p/discover. All uses of discover.Node and discover.NodeID are replaced by their equivalents from p2p/enode. New API is added to retrieve the enode.Node instance of a peer. The behavior of Server.Self with discovery disabled is improved. It now tries much harder to report a working IP address, falling back to 127.0.0.1 if no suitable address can be determined through other means. These changes were needed for tests of other packages later in the series. * p2p/simulations, p2p/testing: port to p2p/enode No surprises here, mostly replacements of discover.Node, discover.NodeID with their new equivalents. The 'interesting' API changes are: - testing.ProtocolSession tracks complete nodes, not just their IDs. - adapters.NodeConfig has a new method to create a complete node. These changes were needed to make swarm tests work. Note that the NodeID change makes the code incompatible with old simulation snapshots. * whisper/whisperv5, whisper/whisperv6: port to p2p/enode This port was easy because whisper uses []byte for node IDs and URL strings in the API. * eth: port to p2p/enode Again, easy to port because eth uses strings for node IDs and doesn't care about node information in any way. * les: port to p2p/enode Apart from replacing discover.NodeID with enode.ID, most changes are in the server pool code. It now deals with complete nodes instead of (Pubkey, IP, Port) triples. The database format is unchanged for now, but we should probably change it to use the node database later. * node: port to p2p/enode This change simply replaces discover.Node and discover.NodeID with their new equivalents. * swarm/network: port to p2p/enode Swarm has its own node address representation, BzzAddr, containing both an overlay address (the hash of a secp256k1 public key) and an underlay address (enode:// URL). There are no changes to the BzzAddr format in this commit, but certain operations such as creating a BzzAddr from a node ID are now impossible because node IDs aren't public keys anymore. Most swarm-related changes in the series remove uses of NewAddrFromNodeID, replacing it with NewAddr which takes a complete node as argument. ToOverlayAddr is removed because we can just use the node ID directly.
2018-09-24 17:59:00 -05:00
if netutil.CheckRelayIP(from.IP, n.IP()) == nil {
p.Nodes = append(p.Nodes, nodeToRPC(n))
}
if len(p.Nodes) == maxNeighbors {
t.send(from, fromID, neighborsPacket, &p)
p.Nodes = p.Nodes[:0]
sent = true
}
}
if len(p.Nodes) > 0 || !sent {
t.send(from, fromID, neighborsPacket, &p)
}
}
func (req *findnode) name() string { return "FINDNODE/v4" }
func (req *neighbors) preverify(t *udp, from *net.UDPAddr, fromID enode.ID, fromKey encPubkey) error {
if expired(req.Expiration) {
return errExpired
}
if !t.handleReply(fromID, from.IP, neighborsPacket, req) {
return errUnsolicitedReply
}
return nil
}
func (req *neighbors) handle(t *udp, from *net.UDPAddr, fromID enode.ID, mac []byte) {
}
func (req *neighbors) name() string { return "NEIGHBORS/v4" }
func expired(ts uint64) bool {
return time.Unix(int64(ts), 0).Before(time.Now())
}